Mojtaba Shirinjani; Siavash Ahmadi; Taraneh Eghlidos; Mohammad Reza Aref
Abstract
Large-scale data collection is challenging in alternative centralized learning as privacy concerns or prohibitive policies may rise. As a solution, Federated Learning (FL) is proposed wherein data owners, called participants, can train a common model collaboratively while their privacy is preserved. ...
Read More
Large-scale data collection is challenging in alternative centralized learning as privacy concerns or prohibitive policies may rise. As a solution, Federated Learning (FL) is proposed wherein data owners, called participants, can train a common model collaboratively while their privacy is preserved. However, recent attacks, namely Membership Inference Attacks (MIA) or Poisoning Attacks (PA), can threaten the privacy and performance in FL systems. This paper develops an innovative Adversarial-Resilient Privacy-preserving Scheme (ARPS) for FL to cope with preceding threats using differential privacy andcryptography. Our experiments display that ARPS can establish a private model with high accuracy outperforming state-of-the-art approaches. To the best of our knowledge, this work is the only scheme providing privacy protection beyond any output models in conjunction with Byzantine resiliency without sacrificing accuracy and efficiency.
Zeinab Salami; Mahmoud Ahmadian Attari; Mohammad Reza Aref; Hoda Jannati
Abstract
Since their introduction, cognitive radio networks, as a new solution to the problem of spectrum scarcity, have received great attention from the research society. An important field in database driven cognitive radio network studies is pivoted on their security issues. A critical issue in this context ...
Read More
Since their introduction, cognitive radio networks, as a new solution to the problem of spectrum scarcity, have received great attention from the research society. An important field in database driven cognitive radio network studies is pivoted on their security issues. A critical issue in this context is user's location privacy, which is potentially under serious threat. The query process by secondary users from the database is one of the points where the problem rises. In this paper, we propose a Privacy Preserving Query Process (PPQP), accordingly. PPQP is a cryptography-based protocol, which takes advantage of properties of some well-known cryptosystems. This method lets secondary users deal in the process of spectrum query without sacrificing their location information. Analytical assessment of PPQP's privacy preservation capability shows that it preserves location privacy for secondary users against different adversaries, with very high probability. Relatively low communicational cost is a significant property of our novel protocol.